Design And Implementation Of CSI Limited Feedback In MIMO-FBMC Communication Systems Based On USRP RIO Platform

In recent years,the process of 5G technology research has been accelerated at home and abroad with the increase for communication needs across all industries.The filter banks(Filter Bank Multicarrier with Offset Quadrature Amplitude Modulation,OQAM-FBMC)based on staggered quadrature amplitude modulation are widely studied by scholars due to their low-frequency spectrum leakage and high spectrum utilization,and it is used as an alternative technology of orthogonal frequency division multiplexing technology in 5G.Meanwhile,MIMO technology makes full use of the space resources and it greatly improves the system capacity without increasing the transmission power.But the precoding in the Frequency Division Duplex(FDD)system requires the feedback of channel state information(Channel State Information,CSI)through the uplink.Since the quantity of CSI feedback raises multiply when the number of antennas increased,the spectrum resources are greatly wasted.Therefore,how to make limited feedback to CSI is of great significance for practical systems.This paper introduces the basic principles of OQAM-FBMC system firstly,analyzes its advantages and disadvantages,and introduces the OQAM-FBMC system for the disadvantages of OFDM.It introduces the OQAM-FBMC system principle and the two different structures of PPN and FS.In the implementation mode,after the performance comparison,the FS structure is finally selected to implement the MIMO-FBMC,and it is illustrated that the MIMO-FBMC can well compensate for the disadvantages of OFDM.Then introduced the key technologies in MIMO,introduced the pilot design method,channel estimation method,channel interpolation,and pre-coding,and then selected the CSI limited feedback technology suitable for practical applications from many methods,summed up The practical application of the CSI limited feedback technology mainly consists of three aspects: time domain,frequency domain,and quantization.After implementation complexity and performance comparison,two methods are selected for implementation in the system.Then we introduced the Universal Software Radio Peripheral(USRP)hardware platform and its supporting software language LabVIEW in detail.This platform can help us to develop the system quickly.The parameter list of the entire MIMO-FBMC system is also described.In this paper,a special pilot for estimating CSI is first designed at the transmitting end,which can solve the interference problem of transmitting pilot data between all antennas and uses the minimum two.Multiplication is used to estimate CSI.After CSI is obtained,subcarriers are clustered to implement limited feedback of CSI in the frequency domain,and linear interpolation is used to complement CSI information at the base station side for a section where subcarriers have null values.On the other hand,this paper also uses the time domain correlation to implement the CSI limited feedback in the time domain,and feedback the impact of periodic intervals on system performance in both dynamic and static cases.The implementation steps of each module in the CSI limited feedback part of the FBMC system are explained in detail.The implementation of this paper is on the USRP hardware platform and both are programmed using LabVIEW language.The programming of LabVIEW FPGA can save a lot of tedious steps for development,more efficient than the traditional FPGA development,and after the realization of CSI limited feedback,the paper finally proceeds.In the test,both the base station and the sender performed video playback.Then,the constellation and SNR values were recorded on the front panel,and the comparison of SNR and feedback overhead between the CSI complete feedback and the cluster-based limited feedback was given.The SNR value is obtained by taking a sufficient number of sample values and averaging.Then,the CSI limited feedback is implemented in the time domain,that is,the CSI feedback period is increased,and the effect of increasing the feedback period on the system performance is tested under both dynamic and static conditions.